68676 Allocating Resources to Control HIV/AIDS in Guangxi Zhuang Autonomous Region: Potential Relevance for China Overview and Synthesis of Main Findings Human Development Unit East Asia and Pacific Region The World Bank August 2007 Document of the World Bank Overview and Synthesis of Main Findings 1. Introduction In recent years, China has dramatically scaled up its response to the HIV/AIDS epidemic. Key steps include the establishment of HIV prevention and control as a priority in the 11th National Five-Year Development Plan (2006-2010), the development of a comprehensive policy framework for HIV/AIDS prevention treatment and care (reflected in the Five-Year action Plan to Control HIV/AIDS 2006-2010), and in March 2006, the issuance of the first legislation directly aimed at controlling HIV/AIDS („Regulations on AIDS prevention and treatment‟, Decree of the State Council No. 457).1 Under this framework, China has implemented a number of innovative harm reduction and prevention interventions and significantly strengthened its epidemiological and behavioral surveillance system. In addition, the government has launched a major treatment, care and support initiative (the “Four Free and One Care�) for poor rural and urban patients. These measures have been accompanied by firm demonstrations of commitment to fighting AIDS by key Chinese leaders. This increased national commitment to fighting HIV/AIDS has been reflected in a rapidly expanding AIDS budget (Figure 1). Central government resources devoted to AIDS doubled between 2003 and 2004, and have increased sixty-fold relative to their level in 2000. In 2006, the central government budget for HIV/AIDS reached 860 million RMB, compared to only 15 million RMB in 2000 (1 RMB=0.125 US$). Local and provincial governments have also increased their contributions to the Figure 1: Central Budget Funding for AIDS HIV/AIDS budget. And there has been a (millions of RMB)) rapid increase in support from the international community, mainly from 1000 the Global Fund, the UN system, and 860 900 801 bilateral agencies. International agencies 800 693 have committed over US$360 million 700 600 (2.9 billion RMB) to China‟s efforts to 500 combat HIV/AIDS for the period 2006- 390 400 2010 (Figure 2).2 300 200 100 100 Despite the large increase in government 100 15 0 resources allocated to HIV/AIDS 2000 2001 2002 2003 2004 2005 2006 activities, surprisingly little is known Year about the use and effectiveness of these funds, especially at the local and provincial level. For example, although the overall magnitude of central government allocations to provinces is known, it is very difficult to trace where and when those resources were spent at the local level; or on what services. It is also difficult to assess the volume and use of provincial and local contributions to HIV/AIDS spending. In this context, the risk of wasting resources is significant. Moreover, there is no 1 For an overview of progress in developing a coherent policy towards HIV/AIDS see Wu, Sullivan et al., 2007. 2 These are planned allocations, which may differ significantly from actual spending. assurance that the funds are reaching those to whom the funds should ultimately be targeted, namely users of HIV prevention and treatment services. Indeed, most existing qualitative and quantitative evidence suggests that patients are financing a disproportionate share of HIV expenditures through Out-of-Pocket (OOP) payments (see CHEI and World Bank, 2007). In this context, there is an urgent need to improve information on actual spending on HIV/AIDS Figure 2: Estimated External Funding for AIDS Programmes in China activities, particularly at the level of provinces and counties, and including at the program level. It Merck Found.; Others; $22,000,000 is also urgent to establish $30,000,000 USG; $58,000,000 Global Fund; mechanisms to evaluate the $168,000,000 impact of said spending. China rapidly needs to develop tools and policies to: DFID; $52,000,000 UN System; $33,880,000 a) assure the right levels and mix of HIV/AIDS prevention, treatment, care and support activities relative to local level Source: UNAIDS. epidemiological and socio- economic conditions; and b) develop monitoring tools and evaluation systems, which will help decision makers at the provincial and local level to evaluate the effectiveness of resource use, assess the performance of implementation agencies and to adjust policies and programs in real time so as to achieve the best value for money. To help the government meet these needs, the Bank designed a program of Analytical and Advisory (AAA) services focused specifically on the Guangxi Zhuang Autonomous Region (thereafter referred to as Guangxi). The AAA acts as a case study with potential relevance for China. Guangxi has the third highest cumulative number of reported HIV infections in all of China, and has been at the forefront in the implementation of harm reduction interventions. The Bank has a long history of collaboration on HIV/AIDS with Guangxi under the Bank Health 9 project. Guangxi health authorities had expressed their interest in and specifically requested Bank technical assistance for capacity building in monitoring and evaluation of AIDS programs and spending. The Guangxi AIDS AAA exercise was programmed to yield the following multiple activities/outputs: 1. The development and testing of a toolkit for tracking public expenditures on AIDS- related services in one high-prevalence county (Liuzhou) in Guangxi province, and accompanying preparation of a report synthesizing and analyzing the expenditure data collected through the above tool (Wang, Zu, Zhao et al., 2007; „A Case Study of AIDS Expenditure Tracking in Guangxi Zhang Autonomous Region, China‟, CHEI-World Bank joint report). 2. The development of tools for cost-effectiveness analysis (CEA) of harm reduction interventions in Guangxi province (including data collection instruments, approach to measuring impact, and a template for costing different types of interventions), and accompanying preparation of a paper on CEA of prevention interventions based on existing programs and data (Masaki, Liu et al., 2007; „Cost-Effectiveness Analysis of Harm Reduction Interventions in Guangxi, China’, Guangxi CDC-World Bank joint report). 3. The preparation of a synthesis report on the epidemiological and behavioral drivers of the epidemic in Guangxi (Prybylski, Guangxi CDC and World Bank, 2006; „Synthesis Report on the HIV and STI Epidemic, Behaviors and Response in Guangxi Province, China’, FHI-Guangxi CDC-World Bank joint report). The value added of the work is derived from the development of innovative methodology, and in demonstrating the utility of rigorous technical analysis in guiding policy analysis. The purpose of this „Overview‟ note is to summarize the main findings of these three outputs. Separate in-depth reports for each of these activities accompany this Overview. 2. Characteristics of the HIV Epidemic in Guangxi Guangxi is situated in southern China and shares a border with Vietnam. It has 14 prefecture- level cities, 7 county-level cities, 57 general counties and 12 ethnic minority autonomous counties in Guangxi, totaling 90 counties/cities. Although Guangxi has traditionally been one of the less developed provinces in China, there has been rapid economic growth in recent years, accompanied by recent expansion of the provincial highway system. In 2005, a new modern highway was opened between the capital city of Nanning and Pingxiang on the China-Vietnam border, which allows for much more efficient transportation to Guangxi‟s largest shipping port, Youyi. Rapid urbanization has accompanied the increase in economic growth. In 2003, 17 percent of families in Guangxi province reported having a family member leave home for more than six months, going to other areas of Guangxi as well as to different provinces. Since the first HIV case was detected in a foreign student in 1989, the HIV epidemic in Guangxi has evolved through three distinct phases. During the first, sporadic period from 1989 to 1995, only 10 cases were reported largely among visitors from other Chinese provinces or abroad. No HIV infections were found among Guangxi residents during this period. During the second epidemic phase, from 1996 to 1997, HIV cases were reported among injecting drug users in Guangxi and among Guangxi residents who had sold blood in China‟s central provinces. During the third phase of rapid growth, the annual number of reported HIV cases in Guangxi increased exponentially: from 530 cases in 1998 to 8,625 in 2005 (Figure 3). Currently, the province has the third highest cumulative number of reported HIV infections in all of China and the third highest estimated cumulative HIV prevalence. However, the cumulative total of 20,604 HIV cases reported through 20053 is a vast underestimate of the actual number of HIV cases in Guangxi province. The Guangxi Working Committee for AIDS Prevention and Control using the UNAIDS/WHO Workbook method2 estimated that there were some 50,000-100,000 persons living with HIV in the province during 2005 alone – which represents a significant share of the estimated 650,000 persons that are living with HIV in all of China (UNAIDS, 2006). 3 See CHEI and World Bank, 2007. This figure is slightly larger than the 19,604 cumulative number of HIV cases reported in the 2006 FHI-WB study. HIV cases have been detected among diverse Figure 3: No. of reported HIV and AIDS cases by year, Guangxi populations including injecting drug users (IDU), 10000 8625 female sex workers (FSWs), 9000 sexually transmitted disease 8000 No. of reported cases 7000 (STD) clinic attendees, 6000 voluntary counseling and HIV cases testing (VCT) clients, 5000 4000 3377 AIDS cases hospital inpatients, sailors, 3000 2132 paid and unpaid blood 1599 1725 2000 832 883 donors and mobile 1000 54 195 530 5352 2 19 56 125 403 populations (both persons 2 2 3 0 coming from other provinces and local 98 00 02 04 6 99 19 20 20 20 -1 residents traveling to other 89 19 provinces). While injecting drug use has been the predominant mode of HIV Source: Guangxi CDC, Prybylski et al, 2006. transmission in Guangxi since 1997, the proportion of reported HIV cases attributable to sexual transmission has also been increasing, from 7.1% in 2002 to 33.3% in 2004. The male-to-female ratio of reported HIV cases has gradually decreased from about 9:1 in the period, 1989-1996 to about 3:1 (22% female) in 2004. The decreasing male-to-female ratio is probably related to increases in HIV infections among sex workers and the sexual partners of HIV-infected men. Despite these evolving patterns, injecting drug use remains the main driver of Guangxi‟s HIV epidemic. Two major drug routes from the „Golden Triangle‟ to Eastern China and international markets traverse the province. According to the Public Security Bureau, by the end of June 2004, there were about 50,000 registered drug users in Guangxi, the fifth largest number in China. Geographic analysis shows that HIV is widely distributed among injecting drug users throughout Guangxi –albeit with significant variations in prevalence rates across locations (Figure 4). Areas with especially high HIV prevalence levels include Baise, Hezhou, Liuzhou (the focus of our case study for both AIDS spending and harm reduction interventions), Qinzhou, and Wuzhou. In Wuzhou, HIV prevalence rose from 0.5% in 1998 to 48.4% in 2003. In Nanning and Qinzhou, HIV prevalence appears to have stabilized at 10-20%. Figure 4: HIV Prevalence Rates among IDUs in Guangxi guilin hechi liuzhou baise hezhou laibin guigang wuzhou nanning chongzuo yulin qinzhou o fangchenggang beihai High >40% Baise Liuzhou Qinzhou Wuzhou Hezhou Middle 15-40% Chongzuo Laibing Nanning Guilin Low <15 Beihai Fangchenggang Guigang Hechi Yulin Although prostitution in China has been illegal since 1949, the commercial sex industry has flourished since the early 1980s. Brothel-based commercial sex is far less visible in China than in other parts of Asia although it still exists and informal transient sex work is far more common and often based out of establishments such as hairdressing salons, hotel bars, dance halls or directly from the street or other public venues. According to the public security agency there were at least 50,000 FSWs in Guangxi in 2001. Long-standing HIV surveillance among sex workers is limited to Nanning and shows that HIV prevalence has stabilized at 10%-13%. Many HIV-infected sex workers also appear to be injecting drug users. However, the crucial nexus between drug injecting and commercial sex risk behavior has been understudied in Guangxi and China in general. HIV remains low among potential clients of sex workers – in at least three studies of over 1,000 long- distance truck drivers in Guangxi, only one has been found to be HIV-positive. However, rates are growing among STD patients, suggesting that sexual transmission of HIV is increasing. Some disturbingly high rates of HIV have been reported among some antenatal samples and further study of this group is clearly required. Men who have sex with men (MSM) are the „blind spot‟ in the understanding of the HIV epidemic in Guangxi - there have been no HIV seroprevalence surveys conducted among MSM in the province. It is clear that Guangxi faces a significant HIV epidemic, which is widely dispersed among injecting drug users throughout the province. The epidemic is highly heterogeneous but also eminently preventable– understanding the diversity of the epidemic between and within prefectures and counties is a prerequisite for informed, prioritized, effective responses. The future size of Guangxi‟s epidemic will depend above all on the scope and effectiveness of HIV prevention programs for high risk groups – especially injecting drug users and their sexual partners, but also programs for sex workers and their clients, and men having sex with men. As illustrated by the accompanying paper on CEA of harm reduction interventions in Guangxi, HIV prevention interventions with injecting drug users and sex workers are relatively inexpensive and provide a high return on investment (Masaki et al, 2007). Needle Exchange Programs are particularly cost-effective, costing only $900 per infection averted when implemented in high risk areas. Coverage is the greatest challenge – high coverage of high impact interventions among populations engaged in high risk behaviors and their sexual partners is essential to reduce HIV transmission and hence avoid a more generalized epidemic. Currently, coverage of Guangxi‟s range of prevention interventions is still far below the desirable threshold. 3. Policy Response to AIDS in Guangxi Political commitment and financing to tackle AIDS in Guangxi have grown as the epidemic has gained severity. Provincial AIDS spending rose from barely 1 million RMB in 1998 (US$ 125,000) to 8 million RMB (US$ 1 million) in 2005. Central government contributions reached 35.84 million RMB (US$ 4.5 million) in 2005.4 Central and government contributions have been complemented by significant local government contributions, but the size of the latter have proven hard to estimate. In the case of Luzhai county (the focus of our detailed AIDS expenditure tracking study), prefecture and county government spending was estimated to account for about 9 percent of total HIV/AIDS financing: this was larger than the share provided by the regional government (7.9%) and almost as large as the central government share (11.6%). Out-of-pocket payments (OOP) by individuals contributed the largest share of total HIV/AIDS spending (61.6%) – see Figure 5. Figure 5: The Sources for HIV/AIDS Financing and their Shares in Luzhai, 2005 International Aids 7.25% OOP 61.61% Central Gov. 11.61% Regional Gov. 7.90% Gov. Finance 28.36% Prefecture Gov. 1.01% County Gov. 7.85% Social Insurance 2.78% Source: CHEI and World Bank, 2007. These resources finance a complex system for providing HIV/AIDS-related services (see Table 1). The designated technical body responsible for HIV/AIDS prevention and control in the Guangxi region is the system of Centers for Disease Control (CDCs) and associated Epidemic Prevention Stations (EPSs). There are 106 CDCs and/or EPSs set up in all 109 counties, with 775 total staff. The system is responsible for core surveillance. Core surveillance components include: HIV sentinel surveillance, behavioral surveillance, and HIV case reporting. The CDC system is also responsible for coordinating HIV VCT and Screening, IEC activities, interventions 4 FHI and WB,2006. among high risk groups, and STI clinical management. AIDS treatment and care is the responsibility of designated hospitals, township health centers and clinics. Table 1: HIV/AIDS Services and Providers Category Provider AIDS treatment and care Inpatient treatment and Care Designated hospitals Outpatient treatment and Care Designated hospitals, township health centers, and clinics HIV Surveillance AIDS and STI reporting All medical facilities, CDC, STD prevention and treatment facilities Sero-prevalence Surveillance CDCs, Institutions for STI prevention and control Behavior surveillance surveys CDCs Ad hoc surveys on HIV/AIDS and CDCs, Institutions for STI prevention and control STIs Prevention IEC activities CDCs, Institutions for STI prevention and control, health care providers5, other sectors, NGOs, social workers/volunteers Interventions among HRGs CDCs, Institutions for STD prevention and control, health care providers, other sectors such as public security, justice, and NGOs STI Clinical Management CDC, Institutions for STD prevention and control, health care providers, pharmacies Others PMTCT MCH centers, designated hospitals HIV VCT CDCs, health care providers HIV Screening CDC, health care providers, blood centers Blood Safety Blood centers, health care providers, NGOs such as Red Cross Associations Condom Promotion CDCs, Institutions for STD prevention and control, health care providers School-based Life Skills Training CDCs/EPS, schools Other Prevention-related activities Program Administration6 Members of the Regional HIV/AIDS Working Committee, and the institutions affiliated to the Committee Source: CHEI and World Bank, 2007. Surveillance Guangxi‟s surveillance system is expanding, and coverage reaches a majority of the urban hubs, some provincial locations, and areas along the main provincial transportation routes. The surveillance system includes national and provincial sentinel surveillance, comprehensive surveillance, provincial behavioral surveillance surveys, and HIV case reporting, however in the case of Luzhai county in 2005, about 5% of the entire HIV/AIDS resources -- around 47,200 RMB or US$5,900 -- were allocated to this surveillance network.7 With just under half of the 5 These include hospitals, MCH centers, Township health centers, village clinics, and other private clinics. 6 It includes policy development, advocacy, program/project development, monitoring and evaluation. 7 The entire HIV/AIDS budget for Luzhai county in 2005 was 963,300 RMB or approximately US$120,400. funds designated for HIV/AIDS and STI surveillance activities, one-quarter for ad-hoc surveys, and the rest divided between behavioral surveys and reporting. Sentinel surveillance, comprised of HIV and syphilis sero-surveillance and a short behavioral questionnaire, is the largest component of the Guangxi HIV surveillance system. In Guangxi, sentinel surveillance has been supported since 2002 by the World Bank Health 9 Project. The target populations of the sentinel surveillance surveys vary across sites and include drug users, sex workers, pregnant woman, STD clinic attendees, TB clinic patients and long distance truck drivers. Although surveillance is extensive, there are significant biases in the major sampling frames used. Sampling of drug users is limited to institutional populations, who may differ significantly from the wider population of drug users. Sex workers are sampled both in institutional settings and the community. Moreover, the phrasing of the questions is inconsistent across populations and many questionnaires do not contain the time references necessary for measuring trends. As limited behavioral indicators are collected within the context of sentinel surveillance, China CDC developed comprehensive surveillance surveys that include a more detailed questionnaire to obtain in-depth behavioral information on high risk populations, as well as a measurement of HIV and syphilis prevalence. However, the sampling framework utilized decreases the capacity for externally representative measurements, which can be replicated in order to measure changes in prevalence and behavioral risks. The surveillance system also includes an intranet-based HIV case reporting system, which is a hugely ambitious undertaking, is state-of-the-art in principle, and has advantages over previous versions, but the system and its users continue to struggle with numerous technical issues. In short, the major strength of the Guangxi surveillance system is its scope. The major limitation is quality - sampling is largely limited to detained populations of drug users and sex workers. Behavioral research is frequently undermined by insufficient privacy, limited expertise in eliciting disclosure of sensitive injecting and sexual practices and insufficient training and support to interviewers. Prevention and Interventions with High Risk Populations Harm reduction interventions targeted at IDUs. The first pilot Methadone Maintenance Treatment (MMT) program in Guangxi was started in Nanning, the capital of Guangxi province, in November 2003. The following year in 2004, three additional pilot MMT projects were initiated in Liuzhou, Hechi, and Wuzhou. Since the beginning of the program, these four facilities have registered over 1,100 IDUs. As shown in Table 2, the number of MMT programs increased rapidly in the period 2004 - 2006, and the number are expected to grow at a much faster rate in coming years. As of May 2006, 11 new MMT sites received approval and are to begin operation later this year. An additional 8 sites have submitted application and are awaiting approval. Table 2: Harm reduction interventions in Guangxi (Number of ongoing programs) ~2001 2002 2003 2004 2005 2006 2007 Total* MMT 1 4 4 (15)** 25 4 NEP 4 5 11 29 31 31 31 SW 2 2 2 24 30 30 (50)** >50 30 Source: Guangxi Center for HIV/AIDS Prevention and Control * as of June 2006 ** the figures are as of June 2006, and the figures in ( ) are the number of sites that are planned by end of 2006 In 1999 the national and local CDCs started the first pilot Needle Exchange Programs (NEP) in Liuzhai and Tiandong, followed by Yongning and Ningming in 2002. An additional 25 sites have been selected for NEP expansion. There are 31 sites currently providing needle exchange services. There are variations in the mix of interventions provided in each NEP site; however, most of the NEP programs in Guangxi follow the general protocols developed by Ministry of Health. A typical needle exchange program includes the harm reduction activities of: (a) collection and safe disposal of used needles and syringes; (b) social marketing of new needles and syringes, including direct distribution and redemption of pharmacy/clinic vouchers; (c) community education; and (d) support of drug use cessation. More than 50,000 IDUs are currently registered with the Guangxi public security office and the actual number of IDUs is estimated to be several times higher. As shown in Figure 6, the growth of the number of harm reduction programs, particularly SW programs and NEPs has lagged the epidemic by several years but expanded rapidly starting in 2003 – 2004. Figure 6: HIV prevalence and harm reduction programs in Guangxi HIV prevalence among IDUs and harm reduction programs in Guangxi 0.6 35 0.5 30 HIV prevalence rate 25 0.4 20 0.3 15 0.2 10 0.1 5 0 0 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 NEP SW MMT Wuzhou Liuzhou Nanning Ningming Pingxiang Condom promotion in entertainment places. Guangxi CDC has carried out the health education and condom promotion in Liuzhou, Pingxinag, Dongxing, Beihai, Hengxian County near Nanning, Hezhou and Yizhou. Among these interventions, condom promotion in Liuzhou has the longest history and possibly the widest population coverage. In these interventions, CDC staff visit entertainment places to distribute IEC material and condoms, offer face to face counseling, demonstrate the right way of using condoms and provide counseling on reproductive health and HIV/AIDS/STD treatment to CSW. VCT services are also offered. Public education & awareness raising. Numerous multi-sectoral HIV awareness campaigns have been conducted among the general population using diverse media. Health education is given to students, mobile populations and others engaging in risky behaviors. Health education, peer education and volunteers have been brought into school lectures. Government departments of public health, public security, judicature, traffic, family planning, the women federation and the labor union have given over 20,000 lectures, printed and distributed more than 13 million IEC materials, and set up more than 28,000 billboards. Treatment & Care The Chinese government has committed itself to providing comprehensive treatment and care to HIV/AIDS patient, under the “Four Free and One Care� program. This includes access to free anti-retroviral treatment for poor patients, free VCT, free PMTCT and free testing for their babies and free education for AIDS orphans. Anti-retroviral treatment and treatment for OIs for about 400 patients is being provided in designated hospitals in Nanning, Liuzhou, Hezhou, and Pingxiang. PMTCT is being carried out in the four national China CARE sites in Pingxiang, Baise, Hezhou and Heng counties. How effective has this policy response been? This study seeks to provide an empirical basis for assessing the effectiveness of the policy response in Guangxi to date. Specifically, we look at: (a) How/where resources are being spent, and in particular are they being spent in line with epidemiological and policy priorities? (b) What is being achieved in exchange for those resources? (c) What has been the impact of spending in the highest priority area – harm reduction interventions with high risk groups? Our main findings are described in the next section. Based on these findings, we then provide some recommendations for improving the use, efficiency and effectiveness of resources devoted to HIV/AIDS in Guangxi and in China, more generally. 4. Main Findings and Recommendations HIV/AIDS Expenditures8 An array of methods was used to study the flow of funds and the use and effectiveness of AIDS spending (Table 3).9 These methods were applied to both province-level data and more in- depth through a case study of HIV/AIDS financing in Luzhai county. 8 This discussion is summarized from CHEI and World Bank, 2007. See full report for details. 9 See CHEI and World Bank 2007, Table 5 and accompanying text for details. Table 3: HIV/AIDS Expenditure Analysis: Methods Used Purpose Method Understanding HIV/AIDS Financing: National Health Accounts; Surveys among -Sources, providers, services, beneficiaries institutions and High Risk Groups; Desk -Equity of HIV/AIDS financing Reviews; Analysis of Kendall Concordance Coefficient Identifying Bottle-neck Issues in Flow of Public Expenditure Tracking, Desk HIV/AIDS Resources Reviews, Interview with Key Informants Collecting Information on the Local HIV Desk Reviews; Delphi Method Epidemiology, Health Seeking Behaviors, etc. The main findings are: 1. As illustrated by the case of Luzhai county, the current structure for HIV/AIDS financing in Guangxi, relies excessively on out of pocket (OOP) spending. OOP expenditures accounted for 61.5 percent of total annual HIV/AIDS spending, representing a large burden on HIV/AIDS patients.10 PHA‟s average annual expenditure for inpatient HIV/AIDS treatment and care alone was 8,100 Yuan in 2005, which was 2.67 times higher than the net average annual income per capita of Luzhai‟s population. This is clearly unsustainable from a patient‟s perspective. Surprisingly, the share of OOP is as high for spending on prevention as it is for spending on treatment – even though much of prevention spending could be considered as contributing to a public good. 11 Government contributions (all levels) amounted to only 28.5 percent of the total HIV/AIDS expenditure. Of this 28.5 percent, more than half were contributions of provincial, prefecture and local governments; while central government contributions represented only 11% of total HIV/AIDS spending. This allocation of responsibility for HIV/AIDS financing differs markedly from what you see in middle income countries, and even in countries with incomes significantly below that of China (Figure 7). Moreover, it seems at odds with the Chinese Government‟s principle that „HIV/AIDS should be financed by different sources, of which government’s input is the prime source’ although it does reflect its maxim that the responsibility is to be shared by different levels of government. 10 In comparison, OOP spending accounted for 57.7% of total expenditures on health. 11 73.2% of all OOP spending went to prevention, and 26.8% to treatment. Blood safety, purchasing clean syringes and condoms, and STI treatment consumed 90% of OOP on HIV/AIDS prevention. Figure 7: Sources of HIV/AIDS Financing in Selected Countries Mexico Chile Costa Rica Panama Venezuela, RB Uruguay Argentina Brazil Thailand El Salvador Guatemala Colombia China (Luzhai) Philippines Paraguay Honduras Bolivia Nicaragua Kenya Laos PDR Ghana Cambodia Burkina Faso Mozambique Rwanda 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Public OOP ROW P Source: CHEI and World Bank, 2007. Estimated form NHA data. Note: ROW = Rest of the world 2. The effectiveness of Government‟s expenditure on HIV/AIDS can be increased significantly by improving both allocative and technical efficiency. The allocation of spending does not correlate well with evidenced-based policy priorities. While spending on prevention activities accounts for the bulk of total spending (69.5%), the distribution of spending across different types of prevention interventions does not reflect the priority that should be given to interventions among high risk groups. Spending on STI prevention and control, PMTCT and M&E are too low, while that on IEC is too high. Given that the epidemic is still largely confined to high risk groups, the current level of spending on IEC activities could be scaled down to leave more resources available for higher priority and more cost effective interventions. As it stands, IEC consumed the largest share of government spending on HIV/AIDS, even though the characteristics of the Guangxi epidemic and cost-effectiveness analysis suggest that the most effective use of resources would be to support low-cost interventions (such as NEP and STI prevention and control) among high risk groups in high prevalence areas. Implementing effective IEC remains a challenge of how to reach at risk populations in an environment that increases their receptivity to preventive messages. The fact that only 0.3 percent of HIV/AIDS expenditure was spent on M&E of ongoing projects suggests that M&E of interventions is not systematic. The CEA of prevention interventions carried out in parallel to the AIDS expenditure work suggests that in practice there are large variations across sites in the costs and efficiency of interventions – much of them linked to the quality of program management and planning. 3. There is significant scope to improve budget preparation and financial management practices at both the central and provincial level. Budget preparation at both central and provincial level is excessively top-down12 and does not reflect overall needs, risk factors, fiscal capacity or implementation capacity at the county and below-county level. Since 2004, a seven step process has been utilized for budget development of centrally earmarked transfers. The Ministry of Health (MOH) first submits the projected size of transfer for the following year to the Ministry of Finance (MOF) in July. In turn, the MOF sets the budget target for HIV/AIDS transfer after balancing various needs and according to the projected revenue in the next year, and sends this figure back to MOH. MOF and MOH then agree upon the budget target before it is sent to the National People‟s Congress for approval the following March. MOH develops the detailed allocation plan by considering the amount available and its annual work priorities. The budget for any given activity is determined by the work volume multiplied by unit costs. However, since no empirical data on unit costs for each activity in China exists, estimations of the different unit costs (by national HIV/AIDS experts) are used instead. MOF reviews the plan and sends its feedback to MOH before MOH revises the plan and re-submits it for approval by MOF. Finally, MOF approves the allocation plan. For regional HIV/AIDS earmarked transfers, the regional government adopted a line-item budget plan, which amount would increase yearly, and with no specification on project outputs. The process was described as “two ups and two downs�. The budget plan for using regional earmarked transfers was usually reviewed and approved by the Regional People‟s Congress in either January or February. Overall, budget preparation was time consuming since adequate tools to facilitate budget development had not been developed and relevant data, such as the unit costs for different activities, were not readily available. 4. Once approved, central earmarked transfers took too long (average of 18 months) to reach the service providers. In 2005, it took 9 months for the central earmarked funds to be released to the provinces. An additional 5 months delay took place between the provincial and the prefecture level, 2 more months between prefecture and county, and 3 additional months within county, before the funds would reach the actual providers. Provincial earmarked funds did not do much better, and there is no evidence of improvement in the timeliness of the flow of funds in the last few years. Another significant bottleneck is that all central and provincial transfers take the form of a „quota‟ for fund allocation not cash. These transfers are only materialized after the county and province have fulfilled their commitments to the required revenue to the provincial and central government respectively. This tends to penalize poor counties, and those with less revenue generating capacity. In practice it means that providers/implementation agencies have to front the funds needed to provide HIV/AIDS services before they can be reimbursed. In the more extreme cases, they will simply not provide services until the transfers have „materialized‟ – often not until the end of the year. 12 Prior to 2006, the central government did not consult provinces when developing the budget plan for earmarked AIDS transfers. This made it difficult to integrate central transfers with the rest of the resources available at the province level. Table 4: Timeliness of Flow of Funds for Central and Regional Earmarked Transfer Milestone Events Central Transfer Regional Transfer 2003 2004 2005 2005 Budget Approval 2003-3 2004-3 2005-3 2005-2 Allocation Plan release 2004-3-22 2004-12-15 2005-12-5 2005-2 Transfer arrival at Region level 2004-3-16 2004-11-25 2005-12-5 Within Region Allocation Plan 2004-10-8 2005-4-22 2006-4-12 Release Transfer arrival at prefecture 2004-10-8 2005-4-19 2006-4-12 2005-4-14 level 2005-11-17 Transfer arrival at County level 2004-10-19 2005-6-9 2006-7-26 2005-4-25 2006-8-23 2005-11-22 Fund from County BoF to BoH 2004-11-23 2005-10-9 2005-9-5 N.A. 2005-11-8 2005-12-7 Fund from BoH to County 2005-7 2005-10-27 2005-9 N.A. CDC, etc. 2005-11-8 2005-12-13 Source: CHEI and World Bank report, 2007. 5. An analysis of the consistency between the size/distribution of central earmarked transfers and provincial needs suggested that the allocation of transfers across provinces is fairly equitable and in correlation with needs. 13 However, the same analysis revealed that the allocation of central, provincial and pooled transfers within the province (e.g. across counties) was not equitable and did not reflect county-level needs and risk factors. The analysis also suggests that share of central and provincial earmarked transfers reaching the county level is too low (only 25% in 2004 and 2005). 6. Current annual budgeting approaches are prone to inflexibility coming from the budgeting principle of „designated fund for designated task‟. This is especially true for transfers from the central government. At the provincial level, conventional „line item‟ budgeting has also suffered from inflexibility and lack of results orientation. Moreover, different timelines and budget methodologies at central and provincial level (and also for donor-supported projects) have meant that there has been no integration of resources from different channels. As a result, unnecessary redundancy and negligence in both project geographic coverage and package of service provision co-exist in Guangxi. In addition, slow disbursement is widespread in both government and donor financed HIV/AIDS projects, mainly due to the fact that a key impediment, low capacity at the grass roots level, was not addressed. 7. Once the money is received, financial management remains a concern. For the management of the central earmarked transfers, the Ministry of Finance and Ministry of Health issued Provisional Methods for the Earmarked Funds for Local Health Development from the Central Government. Some loopholes were identified in its financial management. Proper accounting (i.e. filing of financial documents) at the end users did not take place and supervision was irregular. The regional earmarked transfers suffered the highest financial management risk. Neither a financial management guideline, nor a proper designated accounting system was set up for the regional earmarked transfers. Financial management supervision has not been implemented and to date, an audit has not been commissioned. (Table 5). 13 Although the equity of the allocation may be reduced by the fact that provincial GDP per capita is not taken into account. Table 5: Financial Management Agreements Central Earmarked Fund Regional Earmarked Fund Guidelines Provisional Methods for the No Earmarked Funds for Local Health Development from the Central Government Implementation  No proper filing of  No designated accounting financial documents system  Irregular supervision by the  No supervision Regional Health Bureau  Audit since 2005 Effectiveness Results monitoring  None  None Risk  Moderate risk  High risk Efficiency  Serious delay in fund  Serious delay in fund releasing releasing CEA of Harm Reduction Interventions14 A separate analytical exercise was carried out to estimate the cost and cost-effectiveness of harm reduction interventions in Guangxi Province, so as to help policymakers and program staff mount a well-targeted, cost-effective, evidence-based HIV/AIDS prevention response. This exercise first compiled and analyzed the resources used and their costs for delivering harm reduction services. It then modeled the epidemic impact of behavioral changes produced by these HIV prevention interventions. Recent surveillance data in Guangxi suggests that the majority of new HIV infections are concentrated within high risk groups, such as IDUs and FSWs. For this reason, the study focused on three types of harm reduction activities targeted to high risk groups: methadone maintenance treatment (MMT); needle exchange programs (NEPs); and a program for sex workers. The data were collected from six intervention sites in Guangxi Region during the period of December 2005 to April 2006. The main findings of the CEA report included: 1. Economic costs of MMT services for the year range from US$29,599 to US$73,639, with the unit costs estimated at US$19.0-US$49.4 per client-month, with an average cost of US$33.8 per client-month. Economic costs of needle exchange services range from a low of $9,149 to a high of $19,271 by the different sites, with mean economic costs of $14,210 per year. Unit cost per needle distributed is very low at around US$0.1. The unit cost per condom distributed is $0.45. There are some large variations in prevention costs, outputs and efficiency across programs and services (Table 6). Some of this variation may reflect real differences in the cost and demand structure faced by the respective sites. However, it is also possible that some of this variation in efficiency reflects varying quality of program management and planning. Understanding the determinants of these variations may yield insights that could be translated into higher levels of efficiency. Personnel and recurring services account for a large portion of the total costs, nearly 40% for NEPs and 56% for 14 This summary is extracted from Masaki et al, 2007. See full report for details. MMT programs. Thus, identifying ways of increasing productivity may be a promising avenue for enhancing efficiency. Table 6: Summary Results for the CEA USD Six HIV prevention programs in Guangxi Costs, Outputs and Efficiency Annual economic Site Annual outputs Economic cost per output costs Methadone maintenance programs1 Nanning 73,693 1,493 49.4 per client-month Wuzhou 29,599 1,558 19.0 per client-month Ave, all MMT sites 51,646 1,526 33.8 per client-month Needle exchange programs2 Ningming 9,149 90,773 0.10 per needle/syringe Wuzhou 19,271 157,720 0.12 per needle/syringe Ave, all NEP sites 14,210 124,247 0.11 per needle/syringe Interventions for commercial sex workers3 Pingxiang 30,914 2,515 12.3 per client reached 1. Client-month of services delivered 2. Needles/syringes distributed 3. Clients reached Source: Masaki et al., 2007. 2. The most cost-effective intervention of the three studied here is needle exchange, at approximately $900 per HIV infection averted. The other two interventions both cost about $3,000 per infection averted. These estimates of costs effectiveness are shown in Table 6 below, along with the estimated number of HIV infections averted under three different prevention scenarios (compared to a baseline scenario of no new prevention activities funded). Scenario A is $10 million spent on methadone maintenance, spread over 5 years. Scenario B is $10 million spent on needle exchange, spread over 5 years. Scenario C is $5 million spent on SW programs, spread over 5 years (according to our data, it is not possible with current SW program design to spend all $10 million due to limited numbers of SWs and low unit costs). All estimates are very sensitive to estimates of unit cost and effectiveness. Table 7: Cost and epidemic impact of interventions in high risk areas by different prevention scenarios Methadone Needle Sex worker maintenance Exchange programs A B C HIV infections prevented over 3,722 11,210 1,651 10 years, with 5 years of program operation Cost per infection prevented $2,687 $892 $3,029 (US$) Source: Masaki et al., 2007. Policy Recommendations The different studies carried out under the framework of this AAA program have identified a number of issues that are constraining the effectiveness and impact of HIV/AIDS expenditures in Guangxi province. The findings can directly help Guangxi provincial and local health authorities improve the effectiveness of their local response to HIV/AIDS. Although the findings from these studies cannot be simply extrapolated and generalized, they can illustrate bottlenecks and assist policy makers at central and local levels throughout China identify problems that apply across provinces and localities. While the issues highlighted in this report are specific to Guangxi, they are not limited to Guangxi. Deficiencies in budgetary planning, gaps in surveillance systems, and the cost effectiveness of harm reduction interventions are concerns faced across China. This report can aid policymakers across the country in detecting where the budgetary system is succeeding and breaking down, where gaps in surveillance quality exist, and the cost efficiency of harm reduction interventions given specific profiles of the epidemic. The work further provides specific and generalized recommendations to assist in improving the efficiency, equity and execution in allocating resources to control HIV/AIDS in China. Based upon the study‟s findings and international best practices, we make some general recommendations. 1. Improve the efficiency of public spending on HIV/AIDS in Guangxi by taking action to:  Better align the allocation of spending with the priorities drawn from the epidemiological evidence. Specifically, more spending should be directed to interventions aimed at high risk groups, and to other proven cost-effective interventions such as STI prevention and control, and PMTCT. Spending on general IEC, which currently absorbs the largest share of public spending, can be reduced.  Focus spending on harm reduction in high risk prefectures and cost effective interventions (such as NEP) where the payoff is highest in terms of infections averted.  Increase the share of resources going to local and county level services, as opposed to province level HIV/AIDS programs.  Explore scope for efficiency gains and cost reductions through better program management and planning at the level of intervention sites, improved labor productivity (capacity building and measures to attract more skilled human resources), and savings in use of facilities and staff.  Address capacity constraints at local level (county level and below) and mobilize NGOs and other non-government agents in HIV/AIDS prevention and control.  Improve M&E of inputs and outputs of HIV/AIDS responses. Scale up the collection of basic program operation costs and CEA started under this AAA to cover all intervention sites. Integrate factors such as the coverage of the different services, cost-effectiveness, and total impact in terms of averted new HIV infections into program and budget planning. 2. Improve the equity of public spending on HIV-AIDS by:  Reducing reliance on OOP for spending on HIV services – especially prevention services with public good aspects – by increasing public spending in these areas.  Reduce reliance on OOP for spending on treatment and care through expanded coverage of urban BMI, rural NCMS and MA.  Improve composition of package to be provided for free on the basis of vertical and horizontal equity – include STI screening and treatment especially among pregnant women, treatment and care for OIs, and not just ART.  Realign relative shares of central, provincial and local contributions to better reflect revenue raising abilities and weigh central/provincial contributions by per capita income levels of receiving areas, so that more resources are directed towards poorer counties. 3. Improve budget planning, preparation and execution through measures to:  Increase the coordination of timelines and methodologies for preparation of central and provincial budgets. The central government is already piloting the delegation of budget development for HIV/AIDS prevention and control to the provinces. This commendable move makes synchronization of budgeting for the central and provincial earmarked transfers possible. However, to ensure quality standards of budgeting by the provinces two additional steps are needed: (a) the development of a software tool for budgeting to standardize this practice across provinces; and (b) capacity building among provincial and county staff in budget development.  Upgrade methods for budget development over the medium term to meet international best practices standards. Move from annual budgeting to multi year budget frameworks, and from program and line-item budgeting to results-based budgeting.  Support provincial efforts to develop implementation plans for their five year action plans for HIV/AIDS prevention and control. These plans would be the basis for annual results-based budgeting.  Reduce delays in budget execution, especially in the release of funds from central to provincial level. Consider allocating and releasing central earmarked transfers twice a year so that they would arrive at the local level in a more timely fashion (as has been piloted for the regional earmarked transfers in Guangxi). 4. Improve the foundation of data and knowledge on the effectiveness of HIV/AIDS spending nation-wide by expanding the expenditure tracking and CEA analysis to other localities and high incidence provinces. References (Overview) Wang, S., Q Zhu, Y Zhao, M Tan, Y Zhang, J Liu, Q Wan, H Li, and H Huang, China: AIDS Expenditure Tracking in Guangxi Zhuang Autonomous Region: A Case Study, draft CHEI-World Bank report. E Masaki, X Chen, Q Zhu, Y Chen, M Lu, E Marseille, J Khan, A Revenga, 2007; „Cost- Effectiveness Analysis of Harm Reduction Interventions in Guangxi, China, draft Guangxi CDC-World Bank report. Prybylski (FHI), Guangxi CDC and David Wilson, 2006; „Synthesis Report on the HIV and STI Epidemics, Behaviors and Response in Guangxi Province, China, draft Family Health International-Guangxi CDC-World Bank report. Wu, Z., Sullivan, S., Wang, Y., Rotheram-Borus, M.J., Detels, R., 2007. „Evolution of China‟s response to HIV/AIDS‟, The Lancet, Vol 369 pp. 679-690. (see Chapters 3 to 5 for full set of references).